SERV/MURP: Chrysler’s Space Truck

A schematic diagram of Chrysler’s SERV, an unusual VTOL Shuttle proposal to NASA in 1969-71. This is the final version, which could have mounted a crew capsule on top as shown here, or the capsule could have been swapped out for the MERP D-10, shown in the upper right. Image from the 1971 NASA document Project SERV Final Review. Click for a larger view.

What it was: A contender for the position of Space Shuttle at NASA. Unlike all the other possibilities raised during the Shuttle’s early development, it was a squat single-stage-to-orbit rocket booster that could get to orbit, release a small crew vehicle or cargo container, return to Earth ballistically, and soft-land on its original launch pad under its own power using jet engines. Both sections of the craft were reusable.

Details: A reusable Space Shuttle was originally a small part of a large post-Apollo NASA program, but over the years 1969-1971 it became progressively clearer that it was all they were going to get of that program. As a result the competition to build it was intense, with at least a dozen proposals—solicited and unsolicited—coming to them before they picked the “winged-orbiter-external-tank-two-side-boosters” approach that actually ended up getting built.

NASA had made it known that they were primarily interested in a winged orbiter, and so accordingly almost all of the proposals they received started from that base. But they didn’t explicitly rule out another approach. Chrysler produced the only notable proposal that came from outside the box.

The first iteration of the SERV/MURP came in November 1969, as part of Phase A of the competition. Chrysler had already done a lot of work for NASA, as they were the ones to build the Saturn I-C (the first and largest stage of the Saturn V), and with the Apollo program winding down they were looking to continue using the Michoud Assembly Facility in New Orleans that they’d been using to fabricate it. Their new proposal started from an interesting solution to a basic problem: two of the requirements for the shuttle, cargo capacity and cross-range capability, were at odds. Making a spacecraft better at either one would eat into the other. Chrysler’s idea was to split the two capabilities.

One module (MURP, the Manned Upper-stage Reusable Payload), was a manned orbiter outsourced McDonnell Douglas. It had a high cross-range ability but no real cargo capacity, while the other (SERV, the Single-stage Earth-orbital Reusable Vehicle) would be a single-stage-to-orbit booster with no real ability to move cross-range but with a huge cargo bay. The SERV and MURP would be mated, with the booster lifting the orbiter that was sitting on its tip. The booster would separate once in LEO, deliver its cargo, and then return to Cape Canaveral, while the orbiter could carry its people on their merry way to wherever it was they were headed—be it the proposed NASA Space Station, or to dock with injection stages headed for the Moon.

The MURP was fairly straightforward, a winged craft based on the HL-10 lifting body that NASA had been testing since 1966. There were actually two MURPs in the original proposal, one larger (the D-34) and one smaller (the D-10). The D-34 had 85 cubic meters of internal cargo space, while the smaller had only 5 cubic meters but made up the difference with a cylindrical cargo pod attached to its aft end. As the cylinder was a more efficient use of materials to enclose the space, the D-10 was considerably lighter than its bigger brother, 11,640 kilograms as compared to 16,150. Both would take two crew and carry up to ten passengers, and each had only a small amount of fuel—the SERV would get it into orbit, so it only needed to be able to go to a higher space station orbit on its own and perform a de-orbit burn. It would be covered with a spray-on silicone ablative skin (peeled off and refreshed after every trip) that would protect it from re-entry, and it would be able to land at any landing strip of reasonable length.

While the MURP was interesting, it was essentially just a very small version of what the other contenders for the Shuttle contract were proposing. Where Chrysler diverged most strongly from its competitors was with the SERV.

SERV would have been a tubby rocket, actually wider than it was tall; even with the MURP on the top of it, it would have only been 38.5 meters in height as compared to 28 meters wide.  In contrast, the actually-built Space Shuttle stack was 56.1 meters tall.

It had these measurements because it returned from orbit ballistically, and in fact resembled a gigantic Apollo capsule. This choice let the SERV benefit from the years of aerodynamic research put into that shape. Maximum heat was 1700 Celsius, on the uppermost edge of the blunt underside so it had a heat shield made of a silicone ablative material embedded in a metal honeycomb framework, individual sections of which could be swapped in and out.

The underside of the SERV, with its ablative heat shield and the aerospike engine that would fire the vehicle into orbit. Its concave profile would act as half of an exhaust bell, with the atmosphere as the other. This tuned the rocket’s performance as air pressure changed. Image from Project SERV Final Review. Click for a larger view.

To get into orbit, the SERV used another innovation, an aerospike engine. The typical exhaust  nozzle on a rocket engine is bell-shaped, which has the peculiar effect of making the engine most powerful at one particular air pressure. In other words, a rocket loses efficiency if it’s running at a height lower or higher than the one it’s tuned for—and of course an orbital rocket has to run through a whole range of heights from sea level on up. An aerospike, on the other hand, inverts the bell so that the concave curves are on the outside—now one half of the bell is solid, while the other half is formed by the atmosphere pushing against the side of the exhaust. This has the effect of continuously increasing the size of the imaginary bell as air pressure drops, an on-the-fly reconfiguration that keeps the efficiency of the engine up. On top of that, the SERV engine was going to be extremely powerful: 32 kilonewtons of force, compared to the actual Space Shuttle’s SSMEs at 1.9 kN (of which the Shuttle Orbiters had three), or even the most powerful liquid fuelled engines ever built, the Energiya rocket’s RD-170 at 7.9kN.

As a result it was so powerful that it could get into orbit by itself, even with the dead weight of a MURP attached to the top of it. In all it could lift as much as 39 tonnes to LEO with its single stage, about 40% of what the far more massive Saturn V could do. By the time of the Phase B proposal in 1971, Chrysler tried to make the SERV more attractive by adding to the possible modules that could be mated to its top. As well as a MURP they came up with a ballistic passenger capsule similar in size and shape to an Apollo CM (which was cheaper and smaller than even a D-10, and so the SERV could lift even more cargo), a high fineness ratio nose spike that would make an unmanned version of the craft more aerodynamic so that it could lift even more, and finally a nuclear-rocket launched upper stage that would be perfect for heavy Moon missions or manned trips to Mars.

On its ballistic return to earth the SERV could aim for an area about 15 kilometers in diameter, but unfortunately that was not accurate enough to meet NASA’s requirements. Undaunted, Chrysler proposed to put a ring of 28 jet engines with associated air intake doors around the edge of the SERV’s tubby body, inside its fairing so their profile wouldn’t disturb the booster’s smooth aerodynamics. These would kick in at 7600 meters of altitude, gulping air for oxidizer (and saving some weight by doing so, rather than using liquid oxygen) and push the SERV even closer to its goal. It could even hover for as long as its jet fuel held out. With the aid of the jets, it could get to within 75 meters of its aim point. Two special landing pads would have been built at Cape Canaveral right next to maintenance buildings on the shore of the Banana River so that returning SERVs could be whisked in for a post-mission checkup and refurbishment.

With that kind of performance, Chrysler pointed out in its Phase A proposal that the SERV would be within striking distance of providing a commercially viable suborbital “space airline” between major cities. Almost anywhere on Earth was forty minutes away. The main stumbling block was the cost of fuel, which brought the cost of a ticket to about US$33,000 (in 1969 currency). Chrysler somewhat arbitrarily felt that fuel costs would drop by three-quarters given the volumes that would be made to accommodate NASA’s requirement that any Shuttle would have to fly at a punishing schedule of roughly three times per month. Under those circumstances per-person costs would be only about US$10,900

If given the contract on January 1, 1973, Chrysler anticipated that the SERV/MURP’s first test flight would be towards the end of 1977, with the first operational flight in the first quarter of 1978. The total cost of flying four SERVs with three MURPs was pegged at US$10.01 billion, including operations through the end of 1986.

What happened to make it fail: The SERV, as proposed, had a number of small advantages over the spaceplanes that were submitted to NASA in the Shuttle competition, so it’s a little surprising that it got no traction at all. While it did make it through the Phase A first cut and into Phase B, it was never seriously considered.

The reason for this is speculative as it was the sort of thing that doesn’t get much documentation. Any project at NASA needed a broad constituency, and the SERV failed to make it off the drawing board because its peculiarities turned every possible supporter against it. NASA’s management had come to the conclusion that a large spaceplane was the way to go, and had only allowed other arrangements as a sop to due diligence. NASA’s astronauts didn’t like that the SERV (as ultimately envisioned in the Phase B proposal) could fly unmanned, cargo-only missions. And NASA engineers were concerned about the SSTO/Aerospike engine approach, which is very sensitive to weight and thrust. Marginal failures to meet the proposed engine characteristics or rocket weight can radically reduce the payload that can be carried to LEO, or even keep the craft from reaching orbit at all.

What was necessary for it to succeed: Besides its lack of support, the marginal failures just mentioned were the biggest problem the SERV/MURP had. As proposed, it looks a lot better than the Shuttle that was actually built, but of course the proposals for the actual Shuttle were far more comparable. Besides the aerospike engine, the SERV/MURP needed a number of technologies that were less well worked out than what went into the real Shuttle. If the real Shuttle went over time (it did, by several years) and budget (it did, by about a billion dollars) during development, what would have happened to the SERV, which even in its proposal was going to be more expensive and take longer? The MURP relied on an ablative coating technique that had been a near disaster the one time it was used, on X-15A-2, when that craft was nearly lost due to its ventral stabilizer nearly burning off. The SERV was enormously larger than an Apollo capsule—was it wise to be sure that it would have a similar heating profile?

Ultimately it’s impossible to be sure. It’s easy to take NASA to task for picking a conservative design that ended up being a far more marginal spacecraft than they had hoped, but the fact remains that with similar bad luck on the SERV they could have ended up with something that couldn’t fly at all—and they were being asked to spend the majority of their budget on it for a number of years. As they didn’t have the benefit of hindsight, it becomes easy to see why they went for the more conservative approach.

So to get the SERV to fly you need to make it so that it’s the less radical of the two options. The large winged orbiter of the Space Shuttle had the advantage of years of aerodynamic and thermal studies on other planes, such as ASSET/PRIME, the X-15, the HL-10, and the X-24A. Eliminate that and the orbiter suddenly becomes an unknown, at which point the SERV mated to a ballistic crew capsule (rather than the MURP) starts looking a lot more attractive.